Inspect Demultiplexing

Setup

library(ngsReports)
library(magrittr)
library(scales)
library(pander)
library(tidyverse)

rawFqc <- list.files("../0_rawData/FastQC/", pattern = "zip", full.names = TRUE) %>%
    getFastqcData()
deMuxFqc <- list.files("../1_demux/FastQC/", pattern = "zip", full.names = TRUE) %>%
    getFastqcData()
samples <- read_tsv("../0_rawData/samples.tsv")

Base Qualities

Each of the libraries was inspected for overall quality. Positions 3 & 4 from R2 libraries FCC21WPACXX-CHKPEI13070002_L6 and FCC21WPACXX-CHKPEI13070003_L7 showed clear problems with read qualities. This was likely due to an unsatisfactory nucleotide diversity in the original sequencing run and not enough phiX to overcome this, particularly as all R2 reads will terminate with the restriction site.

Base qualities before demultiplexing.

Read Totals

The first check after demultiplexing is to ensure that read were not assigned to multiple individuals by sabre pe. Read Totals before and after demultiplexing were then checked and the recovery rate was >90% for each library. This process was only applied to the data from the 1996 and 2012 populations as the Turretfield samples were provided after demultiplexing by the collaborators from this dataset. Results indicated that demultiplexing was performed with a high degree of success.

rawReadTotals <- readTotals(rawFqc) %>%
    mutate(Library = str_remove_all(Filename, "_[12].fq.gz")) %>%
    distinct(Library, Total_Sequences)
deMuxReadTotals <- readTotals(deMuxFqc) %>%
    mutate(ID = str_remove_all(Filename, ".[12].fq.gz")) %>%
    distinct(ID, Total_Sequences) %>%
    left_join(samples, by = "ID") %>%
    group_by(Library) %>%
    summarise(DeMultiplexed = sum(Total_Sequences)) %>%
    filter(!is.na(Library))

Recovery rate from demultiplexing the 1996 and 2012 samples

Library	Total Sequences	DeMultiplexed	Recovery Rate
FCC21WPACXX-CHKPEI13070002_L6	69,729,067	63,767,250	91.45%
FCC21WPACXX-CHKPEI13070003_L7	65,481,942	65,031,823	99.31%
FCC229TACXX-CHKPEI13070001_L3	56,981,567	56,588,142	99.31%
FCC2GPDACXX-CHKPEI13070004_L2	78,411,565	78,043,413	99.53%
FCC2GPDACXX-CHKPEI13070005_L3	78,445,394	77,182,810	98.39%
FCC2GPDACXX-CHKPEI13070006_L4	70,874,412	69,781,770	98.46%
FCC2GPDACXX-CHKPEI13070007_L6	66,601,007	66,235,650	99.45%

Read Totals for each 1996 & 2012 samples

Read Totals for each 1996 & 2012 samples

GC Content

GC content for all 1996 and 2012 samples.

Inspection by GC content showed the gc2709 and gc2700 appeared to have an exaggerated peak around 60%, whilst all other samples showed a more broad spread across the range. From the Turretfield samples, pt1125 showed an unexpected peak around 50% indicating that sample may contain reads from a different species. This sample should be excluded from all further analysis.

GC content for all Turretfield samples.

Adapter Content

Adapter content was surprisingly high for some samples, and trimming may be appropriate. Note that this plot understates read lengths.

Adapter Content for all 1996 & 2012 samples.

Read Lengths

Length Distributoin for R1 files

Length Distributoin for R2 files

Sequence Content

Sequence content showing the presence of the RS in both the Turretfield and R2 samples.

Conclusion

• The restriction site should be removed from both R2 and Turretfield samples
• Reads should then be trimmed to ~85bp to deal with Adapter sequences